Data-Driven Design of Novel Polymer Excipients for Pharmaceutical Amorphous Solid Dispersions.

About 90% of active pharmaceutical ingredients (APIs) in the oral drug delivery system pipeline have poor aqueous solubility and low bioavailability. To address this problem, amorphous solid dispersions (ASDs) embed hydrophobic APIs within polymer excipients to prevent drug crystallization, improve solubility, and increase bioavailability. There are a limited number of commercial polymer excipients, and the structure-function relationships which lead to successful ASD formulations are not well-documented. There are, however, certain solid-state ASD characteristics that inform ASD performance. One characteristic shared by successful ASDs is a high glass transition temperature (Tg), which correlates with higher shelf stability and decreased drug crystallization. We aim to identify how polymer features such as side chain geometry, backbone methylation, and hydrophilic-lipophilic balance impact Tg to design copolymers capable of forming high-Tg ASDs. We tested a library of 50 ASD formulations (18 previously studied and 32 newly synthesized) of the model drug probucol with copolymers synthesized through automated photoinduced electron/energy transfer-reversible addition-fragmentation chain-transfer (PET-RAFT) polymerization. A machine learning (ML) algorithm was trained on the Tg data to identify the major factors influencing Tg, including backbone methylation and nonlinear side chain geometry. In both polymer alone and probucol-loaded ASDs, a Random Forest Regressor captured structure-function trends in the data set and accurately predicted Tg with an average R2 > 0.83 across a 10-fold cross validation. This ML model will be used to predict novel copolymers to design ASDs with high Tg, a crucial factor in predicting ASD success.

A Dual Wavelength Polymerization and Bioconjugation Strategy for High Throughput Synthesis of Multivalent Ligands.

Structure-function relationships for multivalent polymer scaffolds are highly complex due to the wide diversity of architectures offered by such macromolecules. Evaluation of this landscape has traditionally been accomplished case-by-case due to the experimental difficulty associated with making these complex conjugates. Here, we introduce a simple dual-wavelength, two-step polymerize and click approach for making combinatorial conjugate libraries. It proceeds by incorporation of a polymerization friendly cyclopropenone-masked dibenzocyclooctyne into the side chain of linear polymers or the α-chain end of star polymers. Polymerizations are performed under visible light using an oxygen tolerant porphyrin-catalyzed photoinduced electron/energy transfer-reversible addition-fragmentation chain-transfer (PET-RAFT) process, after which the deprotection and click reaction is triggered by UV light. Using this approach, we are able to precisely control the valency and position of ligands on a polymer scaffold in a manner conducive to high throughput synthesis.

Purifying Low-Volume Combinatorial Polymer Libraries with Gel Filtration Columns.

Recent advances in oxygen-tolerant controlled/living radical polymer chemistry now enable efficient synthesis of diverse and combinatorial polymer libraries. While library synthesis has been dramatically simplified, equally efficient purification strategies for removal of small-molecule impurities are not yet established in high throughput settings. It is shown that gel filtration columns for chromatography frequently used in the protein science community are well suited for high throughput polymer purification. Using either single-use columns or gel filtration plates, the purification of 32 diverse polymers is demonstrated in a library with >95% removal of small molecule impurities and >85% polymer retention in a single purification step. Doing so replaces the typical procedure of polymer precipitation, which requires solvent optimization for each polymer in a complex library. Overall, this work raises awareness in the polymer science community that gel filtration is amenable to purification of large polymer libraries and can speed up the progress of combinatorial polymer chemistry.

Up in the air: oxygen tolerance in controlled/living radical polymerisation.

The requirement for deoxygenation in controlled/living radical polymerisation (CLRP) places significant limitations on its widespread implementation by necessitating the use of large reaction volumes, sealed reaction vessels as well as requiring access to specialised equipment such as a glove box and/or inert gas source. As a result, in recent years there has been intense interest in developing strategies for overcoming the effects of oxygen inhibition in CLRP and therefore remove the necessity for deoxygenation. In this review, we highlight several strategies for achieving oxygen tolerant CLRP including: "polymerising through" oxygen, enzyme mediated deoxygenation and the continuous regeneration of a redox-active catalyst. In order to provide further clarity to the field, we also establish some basic parameters for evaluating the degree of "oxygen tolerance" that can be achieved using a given oxygen scrubbing strategy. Finally, we propose some applications that could most benefit from the implementation of oxygen tolerant CLRP and provide a perspective on the future direction of this field.

An Oxygen-Tolerant PET-RAFT Polymerization for Screening Structure-Activity Relationships.

The complexity of polymer-protein interactions makes rational design of the best polymer architecture for any given biointerface extremely challenging, and the high throughput synthesis and screening of polymers has emerged as an attractive alternative. A porphyrin-catalysed photoinduced electron/energy transfer-reversible addition-fragmentation chain-transfer (PET-RAFT) polymerisation was adapted to enable high throughput synthesis of complex polymer architectures in dimethyl sulfoxide (DMSO) on low-volume well plates in the presence of air. The polymerisation system shows remarkable oxygen tolerance, and excellent control of functional 3- and 4-arm star polymers. We then apply this method to investigate the effect of polymer structure on protein binding, in this case to the lectin concanavalin A (ConA). Such an approach could be applied to screen the structure-activity relationships for any number of polymer-protein interactions.

Biologically-active laminin-111 fragment that modulates the epithelial-to-mesenchymal transition in embryonic stem cells.

The dynamic interplay between the extracellular matrix and embryonic stem cells (ESCs) constitutes one of the key steps in understanding stem cell differentiation in vitro. Here we report a biologically-active laminin-111 fragment generated by matrix metalloproteinase 2 (MMP2) processing, which is highly up-regulated during differentiation. We show that the ß1-chain-derived fragment interacts via α3ß1-integrins, thereby triggering the down-regulation of MMP2 in mouse and human ESCs. Additionally, the expression of MMP9 and E-cadherin is up-regulated in mouse ESCs--key players in the epithelial-to-mesenchymal transition. We also demonstrate that the fragment acts through the α3ß1-integrin/extracellular matrix metalloproteinase inducer complex. This study reveals a previously unidentified role of laminin-111 in early stem cell differentiation that goes far beyond basement membrane assembly and a mechanism by which an MMP2-cleaved laminin fragment regulates the expression of E-cadherin, MMP2, and MMP9.

Combinatorial Low-Volume Synthesis of Well-Defined Polymers by Enzyme Degassing.

The synthesis of well-defined polymers in a low-volume, combinatorial fashion has long been a goal in polymer chemistry. Here, we report the preparation of a wide range of highly controlled homo and block co-polymers by Enz-RAFT (enzyme-assisted reversible addition-fragmentation chain transfer) polymerization in microtiter plates in the open atmosphere. The addition of 1 µm glucose oxidase (GOx) to water/solvent mixtures enables polymerization reactions to proceed in extremely low volumes (40 µL) and low radical concentrations. This procedure provides excellent control and high conversions across a range of monomer families and molecular weights, thus avoiding the need to purify for screening applications. This simple technique enables combinatorial polymer synthesis in microtiter plates on the benchtop without the need of highly specialized synthesizers and at much lower volumes than is currently possible by any other technique.

Modular and Versatile Spatial Functionalization of Tissue Engineering Scaffolds through Fiber-Initiated Controlled Radical Polymerization.

Native tissues are typically heterogeneous and hierarchically organized, and generating scaffolds that can mimic these properties is critical for tissue engineering applications. By uniquely combining controlled radical polymerization (CRP), end-functionalization of polymers, and advanced electrospinning techniques, a modular and versatile approach is introduced to generate scaffolds with spatially organized functionality. Poly-ε-caprolactone is end functionalized with either a polymerization-initiating group or a cell-binding peptide motif cyclic Arg-Gly-Asp-Ser (cRGDS), and are each sequentially electrospun to produce zonally discrete bilayers within a continuous fiber scaffold. The polymerization-initiating group is then used to graft an antifouling polymer bottlebrush based on poly(ethylene glycol) from the fiber surface using CRP exclusively within one bilayer of the scaffold. The ability to include additional multifunctionality during CRP is showcased by integrating a biotinylated monomer unit into the polymerization step allowing postmodification of the scaffold with streptavidin-coupled moieties. These combined processing techniques result in an effective bilayered and dual-functionality scaffold with a cell-adhesive surface and an opposing antifouling non-cell-adhesive surface in zonally specific regions across the thickness of the scaffold, demonstrated through fluorescent labelling and cell adhesion studies. This modular and versatile approach combines strategies to produce scaffolds with tailorable properties for many applications in tissue engineering and regenerative medicine.

Polymerization amplified detection for nanoparticle-based biosensing.

Efficient signal amplification processes are key to the design of sensitive assays for biomolecule detection. Here, we describe a new assay platform that takes advantage of both polymerization reactions and the aggregation of nanoparticles to amplify signal. In our design, a cascade is set up in which radicals generated by either enzymes or metal ions are polymerized to form polymers that can entangle multiple gold nanoparticles (AuNPs) into aggregates, resulting in a visible color change. Less than 0.05% monomer-to-polymer conversion is required to initiate aggregation, providing high sensitivity toward the radical generating species. Good sensitivity of this assay toward horseradish peroxidase, catalase, and parts per billion concentrations of iron and copper is shown. Incorporation of the oxygen-consuming enzyme glucose oxidase (GOx), enables this assay to be performed in open air conditions at ambient temperature. We anticipate that such a design will provide a useful platform for sensitive detection of a broad range of biomolecules through polymerization-based amplification.

Machine learning in drug delivery.

For decades, drug delivery scientists have been performing trial-and-error experimentation to manually sample parameter spaces and optimize release profiles through rational design. To enable this approach, scientists spend much of their career learning nuanced drug-material interactions that drive system behavior. In relatively simple systems, rational design criteria allow us to fine tune release profiles and enable efficacious therapies. However, as materials and drugs become increasingly sophisticated and their interactions have non-linear and compounding effects, the field is suffering the Curse of Dimensionality which prevents us from comprehending complex structure-function relationships. In the past, we have embraced this complexity by implementing high-throughput screens to increase the probability of finding ideal compositions. However, this brute force method was inefficient and led many to abandon these fishing expeditions. Fortunately, methods in data science including artificial intelligence / machine learning (AI/ML) are providing ideal analytical tools to model this complex data and ascertain quantitative structure-function relationships. In this Oration, I speak to the potential value of data science in drug delivery with particular focus on polymeric delivery systems. Here, I do not suggest that AI/ML will simply replace mechanistic understanding of complex systems. Rather, I propose that AI/ML should be yet another useful tool in the lab to navigate complex parameter spaces. The recent hype around AI/ML is breathtaking and potentially over inflated, but the value of these methods is poised to revolutionize how we perform science. Therefore, I encourage readers to consider adopting these skills and applying data science methods to their own problems. If done successfully, I believe we will all realize a paradigm shift in our approach to drug delivery.

Mapping Biomaterial Complexity by Machine Learning.

Biomaterials often have subtle properties that ultimately drive their bespoke performance. Given this nuanced structure-function behavior, the standard scientific approach of one experiment at a time or design of experiment (DOE) methods is largely inefficient for the discovery of complex biomaterials. More recently, high-throughput experimentation coupled with machine learning methods have matured beyond expert users allowing scientists and engineers from diverse backgrounds to access these powerful data science tools. As a result, we now have the opportunity to strategically utilize all available data from high- throughput experiments to train efficacious models and map the structure-function behavior of biomaterials for their discovery. Herein, we discuss this necessary shift to data-driven determination of structure-function properties of biomaterials as we highlight how machine learning is leveraged in identifying physicochemical cues for biomaterials in tissue engineering, gene delivery, drug delivery, protein stabilization, and antifouling materials. We also discuss data-mining approaches that are coupled with machine learning to map biomaterial functions that reduce the load on experimental approaches for faster biomaterial discovery. Ultimately, harnessing the prowess of machine learning will lead to accelerated discovery and development of optimal biomaterial designs.

AAV6 mediated Gsx1 expression in neural stem progenitor cells promotes neurogenesis and restores locomotor function after contusion spinal cord injury.

Genomic screened homeobox 1 (Gsx1 or Gsh1) is a neurogenic transcription factor required for the generation of excitatory and inhibitory interneurons during spinal cord development. In the adult, lentivirus (LV) mediated Gsx1 expression promotes neural regeneration and functional locomotor recovery in a mouse model of lateral hemisection spinal cord injury (SCI). The LV delivery method is clinically unsafe due to insertional mutations to the host DNA. In addition, the most common clinical case of SCI is contusion/compression. In this study, we identify that adeno-associated virus serotype 6 (AAV6) preferentially infects neural stem/progenitor cells (NSPCs) in the injured spinal cord. Using a rat model of contusion SCI, we demonstrate that AAV6 mediated Gsx1 expression promotes neurogenesis, increases the number of neuroblasts/immature neurons, restores excitatory/inhibitory neuron balance and serotonergic neuronal activity through the lesion core, and promotes locomotor functional recovery. Our findings support that AAV6 preferentially targets NSPCs for gene delivery and confirmed Gsx1 efficacy in clinically relevant rat model of contusion SCI.

A User's Guide to Machine Learning for Polymeric Biomaterials.

The development of novel biomaterials is a challenging process, complicated by a design space with high dimensionality. Requirements for performance in the complex biological environment lead to difficult a priori rational design choices and time-consuming empirical trial-and-error experimentation. Modern data science practices, especially artificial intelligence (AI)/machine learning (ML), offer the promise to help accelerate the identification and testing of next-generation biomaterials. However, it can be a daunting task for biomaterial scientists unfamiliar with modern ML techniques to begin incorporating these useful tools into their development pipeline. This Perspective lays the foundation for a basic understanding of ML while providing a step-by-step guide to new users on how to begin implementing these techniques. A tutorial Python script has been developed walking users through the application of an ML pipeline using data from a real biomaterial design challenge based on group's research. This tutorial provides an opportunity for readers to see and experiment with ML and its syntax in Python. The Google Colab notebook can be easily accessed and copied from the following URL: www.gormleylab.com/MLcolab.

Examining polymer-protein biophysical interactions with small-angle x-ray scattering and quartz crystal microbalance with dissipation.

Polymer-protein hybrids can be deployed to improve protein solubility and stability in denaturing environments. While previous work used robotics and active machine learning to inform new designs, further biophysical information is required to ascertain structure-function behavior. Here, we show the value of tandem small-angle x-ray scattering (SAXS) and quartz crystal microbalance with dissipation (QCMD) experiments to reveal detailed polymer-protein interactions with horseradish peroxidase (HRP) as a test case. Of particular interest was the process of polymer-protein complex formation under thermal stress whereby SAXS monitors formation in solution while QCMD follows these dynamics at an interface. The radius of gyration (Rg ) of the protein as measured by SAXS does not change significantly in the presence of polymer under denaturing conditions, but thickness and dissipation changes were observed in QCMD data. SAXS data with and without thermal stress were utilized to create bead models of the potential complexes and denatured enzyme, and each model fit provided insight into the degree of interactions. Additionally, QCMD data demonstrated that HRP deforms by spreading upon surface adsorption at low concentration as shown by longer adsorption times and smaller frequency shifts. In contrast, thermally stressed and highly inactive HRP had faster adsorption kinetics. The combination of SAXS and QCMD serves as a framework for biophysical characterization of interactions between proteins and polymers which could be useful in designing polymer-protein hybrids.

Polymer Texture Influences Cell Responses in Osteogenic Microparticles.

Introduction: Polymer materials used in medical devices and treatments invariably encounter cellular networks. For the device to succeed in tissue engineering applications, the polymer must promote cellular interactions through adhesion and proliferation. To predict how a polymer will behave in vitro, these material-cell interactions need to be well understood. Methods: To study polymer structure-property relationships, microparticles of four chemically distinct tyrosol-derived poly(ester-arylate) polymers and a commercially available poly(lactic acid-co-glycolic acid) (PLGA) copolymer were prepared and their interactions with cells investigated. Cell loading concentration was optimized and cell adhesion and proliferation evaluated. Particles were also tested for their ability to adsorb bone morphogenetic protein-2 (BMP-2) and differentiate a myoblast cell line towards an osteoblast lineage through BMP-2 loading and release. Results: While cell adhesion was observed on all particles after 24 h of incubation, the highest degree of cell adhesion occurred on polymers with smaller crystallites. At longer incubation times, cells proliferated on all particle formulations, regardless of the differences in polymer properties. High BMP-2 loading was achieved for all particle formulations and all formulations showed a burst release. Even with the burst release, cells cultured on all formulations showed an upregulation in alkaline phosphatase (ALP) activity, a measure of osteoblast differentiation. Conclusions: As with cell adhesion, the polymer with the smaller crystallite showed the most ALP activity. We suggest that smaller crystallites serve as a proxy for topographical roughness to elicit the observed responses from cells. Furthermore, we have drawn a correlation between the polymer crystallite with the hydration potential using surface analysis techniques. Supplementary Information: The online version contains supplementary material available at 10.1007/s12195-022-00729-9.

Martini 3 Coarse-Grained Force Field for Carbohydrates.

The Martini 3 force field is a full reparametrization of the Martini coarse-grained model for biomolecular simulations. Due to the improved interaction balance, it allows for a more accurate description of condensed phase systems. In the present work, we develop a consistent strategy to parametrize carbohydrate molecules accurately within the framework of Martini 3. In particular, we develop a canonical mapping scheme which decomposes arbitrarily large carbohydrates into a limited number of fragments. Bead types for these fragments have been assigned by matching physicochemical properties of mono- and disaccharides. In addition, guidelines for assigning bonds, angles, and dihedrals were developed. These guidelines enable a more accurate description of carbohydrate conformations than in the Martini 2 force field. We show that models obtained with this approach are able to accurately reproduce osmotic pressures of carbohydrate water solutions. Furthermore, we provide evidence that the model differentiates correctly the solubility of the polyglucoses dextran (water-soluble) and cellulose (water insoluble but soluble in ionic liquids). Finally, we demonstrate that the new building blocks can be applied to glycolipids. We show they are able to reproduce membrane properties and induce binding of peripheral membrane proteins. These test cases demonstrate the validity and transferability of our approach.

Machine Learning on a Robotic Platform for the Design of Polymer-Protein Hybrids.

Polymer-protein hybrids are intriguing materials that can bolster protein stability in non-native environments, thereby enhancing their utility in diverse medicinal, commercial, and industrial applications. One stabilization strategy involves designing synthetic random copolymers with compositions attuned to the protein surface, but rational design is complicated by the vast chemical and composition space. Here, a strategy is reported to design protein-stabilizing copolymers based on active machine learning, facilitated by automated material synthesis and characterization platforms. The versatility and robustness of the approach is demonstrated by the successful identification of copolymers that preserve, or even enhance, the activity of three chemically distinct enzymes following exposure to thermal denaturing conditions. Although systematic screening results in mixed success, active learning appropriately identifies unique and effective copolymer chemistries for the stabilization of each enzyme. Overall, this work broadens the capabilities to design fit-for-purpose synthetic copolymers that promote or otherwise manipulate protein activity, with extensions toward the design of robust polymer-protein hybrid materials.

Machine-Assisted Discovery of Chondroitinase ABC Complexes toward Sustained Neural Regeneration.

Among the many molecules that contribute to glial scarring, chondroitin sulfate proteoglycans (CSPGs) are known to be potent inhibitors of neuronal regeneration. Chondroitinase ABC (ChABC), a bacterial lyase, degrades the glycosaminoglycan (GAG) side chains of CSPGs and promotes tissue regeneration. However, ChABC is thermally unstable and loses all activity within a few hours at 37 °C under dilute conditions. To overcome this limitation, the discovery of a diverse set of tailor-made random copolymers that complex and stabilize ChABC at physiological temperature is reported. The copolymer designs, which are based on chain length and composition of the copolymers, are identified using an active machine learning paradigm, which involves iterative copolymer synthesis, testing for ChABC thermostability upon copolymer complexation, Gaussian process regression modeling, and Bayesian optimization. Copolymers are synthesized by automated PET-RAFT and thermostability of ChABC is assessed by retained enzyme activity (REA) after 24 h at 37 °C. Significant improvements in REA in three iterations of active learning are demonstrated while identifying exceptionally high-performing copolymers. Most remarkably, one designed copolymer promotes residual ChABC activity near 30%, even after one week and notably outperforms other common stabilization methods for ChABC. Together, these results highlight a promising pathway toward sustained tissue regeneration.

In vitro and in vivo investigation of a zonal microstructured scaffold for osteochondral defect repair.

Articular cartilage is comprised of zones that vary in architecture, extracellular matrix composition, and mechanical properties. Here, we designed and engineered a porous zonal microstructured scaffold from a single biocompatible polymer (poly [ϵ-caprolactone]) using multiple fabrication strategies: electrospinning, spherical porogen leaching, directional freezing, and melt electrowriting. With this approach we mimicked the zonal structure of articular cartilage and produced a stiffness gradient through the scaffold which aligns with the mechanics of the native tissue. Chondrocyte-seeded scaffolds accumulated extracellular matrix including glycosaminoglycans and collagen II over four weeks in vitro. This prompted us to further study the repair efficacy in a skeletally mature porcine model. Two osteochondral lesions were produced in the trochlear groove of 12 animals and repaired using four treatment conditions: (1) microstructured scaffold, (2) chondrocyte seeded microstructured scaffold, (3) MaioRegen™, and (4) empty defect. After 6 months the defect sites were harvested and analyzed using histology, micro computed tomography, and Raman microspectroscopy mapping. Overall, the scaffolds were retained in the defect space, repair quality was repeatable, and there was clear evidence of osteointegration. The repair quality of the microstructured scaffolds was not superior to the control based on histological scoring; however, the lower score was biased by the lack of histological staining due to the limited degradation of the implant at 6 months. Longer follow up studies (e.g., 1 yr) will be required to fully evaluate the efficacy of the microstructured scaffold. In conclusion, we found consistent scaffold retention, osteointegration, and prolonged degradation of the microstructured scaffold, which we propose may have beneficial effects for the long-term repair of osteochondral defects.

Machine learning in combinatorial polymer chemistry.

The design of new functional polymers depends on the successful navigation of their structure-function landscapes. Advances in combinatorial polymer chemistry and machine learning provide exciting opportunities for the engineering of fit-for-purpose polymeric materials.